Berhampur 70918*19311 CALL GIRLS IN ESCORT SERVICE WE ARE PROVIDING
McKinsey on semiconductors
1. Number 1,
Autumn 2011
McKinseyon Semiconductors
McKinseyonSemiconductorsNumber1,Autumn2011
Will analog be as
good tomorrow as it
was yesterday?
Getting Mo(o)re out of
semiconductor R&D
LED at the crossroads:
Scenic route or
expressway?
Mastering variability
in complex
environments
50 57 64 71
From source
to drain: Fixing the
supply chain
42
Getting customers
to say, ‘The price
is right!’
15
Creating value in
the semiconductor
industry
5
The challenge
of China
The evolution
of business models in a
disrupted value chain
22 33
3. 1
5
Creating value in
the semiconductor
industry
In light of increasing
consolidation throughout
the semiconductor
value chain, companies
that wish to succeed
must move quickly to
close capability gaps.
2
Introduction
57
Getting Mo(o)re out of
semiconductor R&D
Excellence in the R&D
function is only one piece
of a world-class product-
development strategy. Via
a new diagnostic and
jointly developed plan,
we can significantly
reduce time to market for
new chips while also
improving overall quality.
42
From source to drain:
Fixing the supply chain
Supply-chain excellence
has proved elusive
for semiconductor players,
but a handful of
initiatives can yield
significant improvement.
15
Getting customers
to say, ‘The price
is right!’
In a world where too many
companies still believe
that ‘if we build it they will
come,’ practitioners of
semiconductor sales know
otherwise. This article
lays out the elements
essential to extracting the
most value for current
products by effectively
communicating the true
value of those products
to customers.
33
The evolution of
business models in a
disrupted value chain
The progress predicted
by Moore’s Law has
slowed in recent years.
Players across the
semiconductor value chain
must adjust their
approaches to compete as
the industry continues
to evolve.
71
Mastering variability
in complex
environments
Variability adds cost to
semiconductor production
systems, but the ability
to cope with it can also be
a critical source of profit.
A new approach goes
beyond traditional tools
to help companies
control variability in their
processes and make
intelligent trade-offs in
order to maximize return.
22
The challenge
of China
As barriers to Chinese
competition weaken,
local and foreign semi-
conductor players
must consider issues such
as intellectual property
and knowledge transfer
to fully capture
opportunities in this
important market.
64
LED at the crossroads:
Scenic route or
expressway?
Although adoption of
LED lighting has been slow,
roadblocks can be
overcome with a compre-
hensive approach
that includes operational
improvements, better
marketing of products, and
other efforts.
50
Will analog be as
good tomorrow as it
was yesterday?
Many worry that 300mm
manufacturing capacity
will destabilize pricing
across the analog semicon-
ductor market. We
argue that only a few seg-
ments have reason to
be concerned.
McKinsey on Semiconductors
Number 1, Autumn 2011
4. 2
Introduction Given the enormous and endlessly expanding role
played by technology in the world economy, the
layperson might be forgiven for assuming these are
easy times for semiconductor companies. As our
inaugural edition of McKinsey on Semiconductors
notes, things are rather more complicated.
On the one hand, the semiconductor industry is
certainly enormous: it has grown to almost
$300 billion, driven by a cycle of continuous
improvements in technology, growth in demand,
and innovation in end-use applications. On
the other hand, as the authors of “Creating value in
the semiconductor industry” note, while the
semiconductor industry contributes dispropor-
tionately to growth in US labor productivity
and delivers tremendous value to consumers, most
chip makers capture only a small percentage
of the value they help create. In fact, excluding
Intel, which made handsome profits indeed,
the industry destroyed approximately $47 billion
in value from 1996 to 2009. If semiconductor
players are to meet market pressure to grow, they
must lead convincingly on technology in their
segment or grow in subsegments where they can
differentiate themselves.
China is a critical source of growth for nearly all
established semiconductor players. Luckily
for them, China has not yet developed a strong
André Andonian,
Harald Bauer, Sri Kaza,
Ulrich Naeher,
and Nick Santhanam
5. 3
indigenous semiconductor business, despite
20 years of effort. However, the forces that have
held China back are weakening, and the
government has launched a slate of initiatives
aimed at developing strengths in three
important new semiconductor markets: cloud
computing, the “Internet of Things,” and
electric vehicles. The result, as the authors of “The
challenge of China” show, is that China will
fight to hold on to its own market and may even
compete for the developed-world market, too.
Many of the industry’s challenges relate to
keeping up with rapacious demand for new and
better products. So it is ironic that “LED at
the crossroads: Scenic route or expressway?”
concerns an undoubted technological
advance that is making extraordinarily slow
progress into general use. The authors
discuss five barriers to adoption that, if addressed,
could lead to an LED-dominated lighting
marketplace by 2015.
Indeed, one might say the whole industry is at a
crossroads: breakthrough innovations are needed
to advance to the next node. Costs for manu-
facturing equipment are rising exponentially. Few
can afford to compete on the technology
frontier created by Moore’s Law. And growth is
concentrating only in select segments such
as smart phones and mobile computing. Of
course, the challenges differ by subsegment. “The
evolution of business models in a disrupted
value chain” proposes models for success in the
fabless segment, explores the limits of vertical
disintegration, and discusses how miniaturization
is bringing success to outsourced semiconductor
assembly and test (OSAT) players.
One piece of the industry has been free from
Moore’s Law’s punishing investment implications:
the analog segment has historically been
stable and profitable. In “Will analog be as good
tomorrow as it was yesterday?” the authors
examine the implications of the industry’s first
300mm-based manufacturing capacity.
They ask whether this stability and profitability
will endure.
Despite many unsettling strategic issues, all
semiconductor companies must still find ways to
improve bottom-line results in the near term.
The supply chain offers one important opportunity.
Obviously, an industry that combines high
capital intensity with long cycle times and a
position far down the value chain will suffer supply-
chain difficulties. But some are avoidable.
As the authors of “From source to drain: Fixing
the supply chain” note, supply-chain performance
varies significantly, even within the same
6. 4 McKinsey on Semiconductors Autumn 2011
applications. The authors analyze excellent supply
chains and discuss what it takes to create one.
Pricing represents another opportunity. The
semiconductor industry, in which price declines are
as certain as death and taxes, is particularly
prone to discounting to maintain market share. The
authors of “Getting customers to say, ‘The price
is right!’” explain how companies can school their
sales forces to get customers to see products
through the lens of their value, rather than their
cost. Only through value selling, they suggest, can
the discounting cycle be broken.
Next, we take on research and development.
How can companies overcome a long history of
time and budget overruns in the face of ever-
changing customer expectations? The authors of
“Getting Mo(o)re out of semiconductor R&D”
explain that getting R&D right implicates many
functions, including marketing, sales, pro-
duction, and even supply chain, and so urge a
broader view of the product-development
process, in which all involved functions are
engaged. By examining one or two end-
to-end cycles in, say, a new feature, a team can
observe and address all the touchpoints,
loops, interfaces, and delays throughout the organ-
ization, reducing time to market by as much as
30 percent without compromising quality.
We close this issue with “Mastering variability in
complex environments.” The combination
of semiconductors’ complexity and the pace of
industry change make variability—a deep
challenge to profitability—very difficult to manage.
What can be done? The authors discuss a way
of quantifying the impact of changes in lead time
and utilization on variability that can help
manufacturers manage that variability beyond what
can be done with traditional methods.
We hope these essays encourage you on your
journey toward excellence. We invite comments at
McKinsey_on_Semiconductors@McKinsey.com.
Harald Bauer
Principal
Ulrich Naeher
Director
Nick Santhanam
Principal
André Andonian
Director
Sri Kaza
Associate principal
7. 5
Despite its moderate size, the semiconductor
industry contributes disproportionately to growth
in US labor productivity and delivers tremen-
dous value to consumers. The industry, along with
the electronics industry it does so much to
power, contributed more than 25 percent of total
US productivity growth from 1995 to 1999—
more than any other sector. That four-year period
outshined overall productivity growth from
1987 to 1995, according to an analysis published
by the McKinsey Global Institute.1
Much of the tremendous growth seen in the
electronics industry over the last three decades
comes directly from the increasing power and
decreasing price of semiconductors, a function of
Stefan Heck,
Sri Kaza,
and Dickon Pinner
Creating value in the
semiconductor industry
Moore’s Law.2 This performance improvement
enables the electronics industry to continually
produce devices and systems that are smaller, more
powerful, and richer in features at lower prices.
It has famously been noted that if the automotive
industry had achieved similar improvements
in performance in the last 30 years, a Rolls-Royce
would cost only $40 and could circle the globe
eight times on one gallon of gas—with a top speed
of 2.4 million miles per hour.
However, most chip makers capture only a small
percentage of the tremendous value they create;
consumers receive the lion’s share. Indeed, despite
its large positive impact on overall economic
growth, the semiconductor industry (excluding
In light of increasing consolidation throughout the semiconductor value chain,
companies that wish to succeed must move quickly to close capability gaps.
1
US productivity growth
1995–2000: Understanding
the contributions of infor-
mation technology relative to
other factors, McKinsey
Global Institute, October 2001.
2
According to Gordon Moore,
a founder of Intel, the number
of transistors that can be
fitted into a single chip doubles
roughly every two years,
resulting in both faster perfor-
mance and lower cost.
8. 6 McKinsey on Semiconductors Autumn 2011
Intel) destroyed approximately $47 billion
in value for shareholders between 1996 and 2009
(Exhibit 1). To put that figure, and the signifi-
cant disparity seen in the industry, into context,
Intel alone created about $57 billion in value
during that same time period.
The economic challenges that the semiconductor
industry faces can be attributed to a confluence of
two factors: cyclicality, and rising costs in
RD and on the capital-investment side of the
ledger, due to the increasing costs of upgrading
existing fabrication plants and building
new ones.
The cycle, while bad for the industry, is in
some ways a blessing for underperformers, who
have been able to stay in business because
the profits they generate during a cyclical upturn
enable them to sustain their operations during
a downturn and attract funds for capital invest-
ments beyond market requirements, which
initiates the next cyclical downturn. Government
interest in building semiconductor industries—
most recently in China and India—accentuates
this problem.
As for RD, chip makers invest heavily, driven to
meet the expectations of Moore’s Law: costs have
Exhibit 1
Positive economic profit (EP)1
$ billion
Negative EP
The semiconductor industry, excluding Intel, destroyed
$47 billion of value from 1996 to 2009.
MoSC 2011
Value creation
Exhibit 1 of 7
1Positive EP in each year of the time period. In addition, Intel had a positive EP of $57 billion during this
period. EP is calculated as net operating profit less adjusted taxes – (capital charge, where capital charge is
invested capital at previous year end × weighted average cost of capital).
Source: Corporate Performance Center Semiconductor Database; McKinsey analysis
91 –47
91
–138
TSMC 14.3
Samsung 14.0
Qualcomm 13.6
Texas Instruments 9.5
Applied Materials 6.0
Mediatek 5.1
Linear 3.1
Maxim 2.4
Rohm 2.1
Analog Devices 2.1
Altera 1.6
Xilinx 1.4
KLA 1.4
Nvidia 1.4
Microchip 1.2
Synopsys 1.2
ASML 1.0
Others 10.0
9. 7Creating value in the semiconductor industry
naturally risen along with the ever-increasing
complexity of the chips. In addition, the
investment hurdle for building a state-of-the-art
chip fab continues to rise.
All that said, it is important to remember that
the $47 billion of destroyed value is an aggregate
figure made up of many losers and several
disproportionate successes. Indeed, in many
segments, the top performer generates more than
100 percent of the total value. How do the top
performers succeed? They implement operational-
improvement programs for product lines that can
hit acceptable targets for return on invested capital
(ROIC), and judiciously divest those that cannot.
Companies that wish to thrive must follow this
example. They must optimize for ROIC rather than
share or gross margin, a process that entails
identifying improvement levers relating to each com-
ponent of ROIC and designing initiatives targeted
to each. Lean operations approaches, including best-
practice manufacturing techniques, exert direct
impact on ROIC and are therefore key levers in this
first step.
The companies that have successfully followed
this two-step model have achieved improvements in
ROIC in the range of 5 percentage points. Some
companies have improved ROIC by as much as 20 to
30 percentage points.
Understanding the sources of
value destruction
Although an analysis of income statements shows
a number of profitable players in the semiconductor
industry, most players are not able to generate
economic profit; that is, their ROIC lags behind
10. 8 McKinsey on Semiconductors Autumn 2011
their weighted average cost of capital. Indeed,
disaggregating the industry by business model and
subsegment reveals that in most segments, only
one or two players create value.
As we have indicated, the industry as a whole
has struggled to generate economic profit because
three factors present unique challenges to
chip manufacturers.
Historically, the semiconductor industry
has shown strong cyclical behavior. During
a typical upturn of one to two years, most
companies generate profits, which they use to
sustain their operations during the down-
turn. In addition, many players use their strong
performance during an upturn to entice
investors in the public markets or get new
loans to fund capital investments; in many cases,
governments subsidize these refinancings
(Exhibit 2).
But precisely because investment runs ahead of
market demand in the upturn, the period
is followed by a longer downturn or a very slow
growth period, during which poor per-
formers struggle. There is some evidence to
suggest that both the amplitude and time
frame of the industry’s cyclicality is moderating,
but it is likely that some degree of cyclicality
will remain.
The skyrocketing costs of RD and the increasing
amount of capital required to build a state-of-
Exhibit 2
Sources of financing for poor performers1
1996
110
100
90
80
70
60
50
40
30
20
10
0
–10
–20
–30
1998 2000 2002 2004 2006 2008 2010
Financing continues to flow from private investors and
governments as poor performers fail to deliver.
MoSC 2011
Value creation
Exhibit 2 of 7
1This includes the 59 players with the lowest average economic profit/revenue.
Source: Compustat; Corporate Performance Center analysis
Cumulative operating
cash flow net capital expenditure
Net debt financing
Equity financing
net of dividends
and buybacks
Performance improvement
is contingent on change in
economic behavior
• Investors and governments
cease to fund value-destroying
players in the future
• Players focus not only
on the top line but also on
operational efficiency
$billion,annual
11. 9Creating value in the semiconductor industry
the-art fab add to the industry’s economic
challenges. Chip makers continue to pour money
into RD as new designs and process tech-
nologies become increasingly expensive to develop.
In 2009, RD spending amounted to approx-
imately 17 percent of industry revenue for semi-
conductor companies (up from 14 percent
a decade earlier) versus 3 percent for automakers,
to take one example. The cost of building
leading-edge fabs continues to increase as well;
for example, the average 8-inch fab costs
$1.6 billion to build, while a state-of-the-art
12-inch fab costs $3 billion to $4 billion. Similarly,
the costs for developing process technologies on
new nodes is increasing dramatically; for example,
the average cost of developing a 90-nanometer
logic process technology is approximately
$300 million, while the cost of developing a modern
45-nanometer logic process technology
is approximately $600 million, representing a
doubling of spend in roughly five years
(Exhibit 3).
Exhibit 3
Integrated device manufacturers: company revenue required to
sustain leading-edge fab investments
Only Intel
and Samsung
can invest
beyond 22nm
2009 semiconductor revenue, $ billion
Intel
2.9
90nm
2004
65nm
2006
45nm
2008
90nm
2004
65nm
2006
45nm
2008
22nm
2012
22nm
2012
3.8
5.2
7.0 9.5
Sony
ST
NXP
Samsung
Freescale
Infineon
Texas
Instruments
Leading-edge fab investment Process-development cost
Fewer logic integrated device manufacturers are able to
sustainably invest in leading-edge nodes.
MoSC 2011
Value creation
Exhibit 3 of 7
1Nanometers.
Source: iSuppli; literature search; McKinsey analysis
Minimum-revenue-
level assumptions
• RD per product (fab
investment/process
development, $ billion):
– 65nm:1 2.4/0.6
– 45nm: 3.3/0.7
– 32nm: 4.4/1.0
– 22nm: 6.0/1.4
• Node developed
over 3 years
• RD as 10–12% of
revenue
• Node introduced every
2 years
17.5
9.7
8.5
4.5
4.5
3.4
32.4
3.2
2.4
2.6
3.3
6.5
32.4
17.5
9.7
8.5
4.5
4.5
3.4
3.2
Revenue level required
for sustainable
leading-edge fab
investment
12. 10 McKinsey on Semiconductors Autumn 2011
Exhibit 4
Capital spending/semiconductor
revenue is decreasing . . .
Criticalexposuresystemcost,
$thousand
Line width, nanometers
. . . even though capital
costs are rising
%
1995
35
30
25
20
15
10
5
0
100,000
G-line
I-line
KrF
ArF
ArF-1
EUV
10,000
10,000
1,000
150mm
200mm
300mm
450mm
1,000
100
100 02000 2005 2010 2015
‘Fab lite’ strategies have reduced capital expenditures, and at
the same time, overall capital costs are rising sharply.
MoSC 2011
Value creation
Exhibit 4 of 7
Source: IC Insights; IC Knowledge
In response to these higher costs, many
semiconductor companies have resorted to “fab
lite” strategies, outsourcing an increasingly
large fraction of their chip production to dedicated
manufacturing foundries. Although this has
resulted in an overall net reduction of capital
expenditures in the industry, from an average of
approximately 27 percent of revenues (from
1996 to 2001) to approximately 20 percent of
revenues (from 2002 to 2009), it has also
led to intense cost pressure on chip makers that
continue to handle all their manufacturing
in-house (Exhibit 4). The shift of manufacturing to
Asia has created additional cost pressures
on those that have yet to transfer operations to
lower-cost locations.
Prices also remain under pressure in the industry
as consumer applications become the main
force driving the semiconductor market. The much
higher elasticity of demand as prices decline
has further accelerated the erosion of average
selling prices.
All these pressures are intensified by the shift in
the end-user market to Asia. Furthermore, the lack
of a “killer app” on the horizon—and the slower
growth of traditional large, high-growth markets
such as PCs and mobile phones—means that
the economic pressures on the industry are not
likely to abate anytime soon.
Learning from the top performers
A handful of semiconductor players have
consistently generated a disproportionate amount
of value in this industry. An analysis of the key
attributes of these companies, as well as those of
the leading players in other industries, sug-
gests the two major lessons noted earlier for those
who seek to capture economic profits in
semiconductors: successful players work to
improve ROIC where it can be satisfactorily
13. 11Creating value in the semiconductor industry
improved, and they aggressively prune product
portfolios of businesses that do not look likely to
become sufficiently profitable.
As far as ROIC is concerned, top performers
focus on changing the dynamics and structure
within a given segment as they seek to build
leading positions early on. Acquiring and holding
a market share of 40 percent or more within
a segment enables companies to drive higher
profits (Exhibit 5). Such companies typically have
closer relationships with key customers,
advanced RD processes that yield better innova-
tion road maps (which are also more closely
aligned with the key value drivers for their segment),
deeper insight derived from having a more
complete picture of where the market is going, and
in many cases, a greater ability to maintain
margins through downturns.
To achieve this kind of performance, semiconductor
companies must optimize ROIC by executing
operational-improvement programs, including
but not limited to making lean operational
improvements, targeting profitability (rather than
other measures), improving asset utilization,
and tuning their capital-asset strategy (that is,
make versus buy) to further improve return
on capital. To target areas for improvement, a
detailed ROIC tree can be used to disaggregate the
components of revenue, cost, and invested
capital and thus identify the main value-creation
levers for each component. Exhibit 6 lists
examples of value-creation levers and the impact
that these levers help companies achieve.
By helping companies implement lean-
manufacturing techniques, we have assisted
more than 10 semiconductor companies
in increasing the throughput of their fabs by 20 to
30 percent (with minimal additional capital
expenditure). Naturally, this has been a significant
driver of improved ROIC, as well as incremental
gross margin. These gains have been achieved by
Exhibit 5 Averagereturnoninvestedcapital,
2007–09,%
Market share, 2009
Lattice
Freescale (DSP)
AMD (MPU)
Hynix (DRAM)
Samsung (DRAM)
Applied Materials
(deposition)
UMC
TEL
TSMC (foundry)
Xilinx (PLD) Intel (MPU)
Texas
Instruments
(DSP)
0
50
40
30
20
10
0
–10
–20
–30
20 40 60 80 100
A high segment market share enables companies to shape
their future and earn higher returns.
MoSC 2011
Value creation
Exhibit 5 of 7
Source: Corporate Performance Center Semiconductor Database; iSuppli; Gartner
Leading player
No. 2 or 3 player
14. 12 McKinsey on Semiconductors Autumn 2011
Exhibit 6
Key levers to drive ROIC
Companies have employed a number of strategies
to achieve impact.
MoSC 2011
Value creation
Exhibit 6 of 7
1Return on invested capital.
Revenue
CostROIC1
Capital
Strategic
Tactical
Fixed
Variable
• Develop products and drive efficiency and effectiveness
in new-product introduction
• Employ a new market-development strategy
• Pricing: improve discipline and value communication
• Sales-force effectiveness: optimize coverage
and productivity of field sales force
• Account profitability: realize value from investing
in key accounts
• Improve fab throughput and drive lean operations
• Employ strategic sourcing focused on total cost of materials
• Optimize consumption to reduce material usage
• Optimize supply chain
• Use fabless/“fab lite” strategy and optimize footprint
• Improve capital-expenditure planning and purchasing strategy
• Reduce general and administrative costs,
optimize performance
• Optimize RD spend/portfolio, rationalize products
–
÷
+
+
maximizing overall equipment effectiveness,
a technique that exposes all the losses attributable
to bottleneck machines in a 24-hour period,
thereby allowing companies to focus on reducing
the largest losses. This technique was as
effective in 4-inch, 5-inch, 6-inch, and 8-inch fabs
(the older, trailing-edge fabs) as it was when
deployed in leading-edge 12-inch fabs.
In trailing-edge fabs, most of the improvements
are captured from increasing the uptime of
bottleneck machines, for example, by minimizing
machine changeovers and setups and optimizing
material handling to ensure that a bottleneck
machine is never left idle. By contrast, in leading-
edge fabs, many of the improvements come
from reducing the process time of an individual
wafer by tailoring the sequence of tasks of
the bottleneck machine to a specific “recipe” (the
unique flow of manufacturing process steps
required to fabricate the wafer) and eliminating
recipe redundancy. For example, dielectric
thin-film deposition times can be decreased, with a
corresponding increase in the throughput of
deposition equipment, by reducing the thickness
of excess dielectric material. This has the
added benefits of increasing both the throughput
of chemical-mechanical-planarization (CMP)
machines (because less excess material is removed
in the polishing process) and the lifetime of the
CMP pads.
Another lever that can help improve ROIC is
pricing, and we recommend chip makers use value-
based pricing and transactional pricing to
drive revenue increases of 2 to 7 percent. Value-
based pricing processes enable companies
to set prices equivalent to the value perceived by
15. 13Creating value in the semiconductor industry
customers by identifying the individual value
drivers of a product, interviewing customers to
understand the importance of each of these drivers
to their purchasing decisions, understanding
the degree of differentiation the company possesses
with regard to each driver, and translating this
value into price. Transactional pricing, by contrast,
focuses on minimizing the leakage of value in
the final price relative to the list price. This leakage
is analyzed with regard to variance (differences
in discounting or margin performance), slippage
(deviations from established policies, guide-
lines, or programs), and structure (suboptimal
pricing structures, processes, or delegation
levels, resulting in unnecessarily low net prices).
Setting aside ROIC, the second main lever involves
proactively managing product portfolios: investing
in market segments that are growing, either
organically or through acquisition, and divesting
segments in which growth or margins are low.
In reviewing its portfolio, a company may find that
it includes some fast-growing businesses with
high profit margins as well as other businesses in
which the company has achieved limited suc-
cess despite years of investment. Top-performing
companies actively evolve their portfolios as
markets mature or become less attractive. Rather
than engaging in a price war to increase their
share of a stagnating market, for example, they
drop out of businesses that offer little hope
of profitability (Exhibit 7).
Several top performers have been particularly
successful with this approach. Texas Instruments
has divested more than 15 lower-growth, lower-
Exhibit 7
Choice of market is the most important
contributor to growth . . .
Sources of growth
Choice of market/
market growth
Market-share gain
MA
Contribution to growth1
Average contribution for semiconductor
peer group,2 2005–08, %
. . . and companies’ performance in
choosing markets differs widely
It has become even more critical for semiconductor companies
to focus on the right markets.
Growth from choice of market
Yearly growth attributable to choice
of market, 2005–08, %
MoSC 2011
Value creation
Exhibit 7 of 7
1Only positive contributions to growth have been included in the analysis.
2AMD, Broadcom, Infineon, Intel, Mediatek, NEC, NXP, Panasonic, Qualcomm, Sony, ST, Texas Instruments, and Toshiba.
Source: Annual reports; McKinsey analysis of granularity of growth
70
19
11
Top performer
9.2
Worst performer
1.6
5×
Companies’ ability to identify the right
markets to compete in has a
significant influence on their total
growth performance
17. 15
Gaurav Batra and
Olivia Nottebohm
Getting customers to say,
‘The price is right!’
In a world where too many companies still believe that ‘if we build it they will come,’
practitioners of semiconductor sales know otherwise. This article lays out the
elements essential to extracting the most value for current products by effectively
communicating the true value of those products to customers.
Keeping the marketing and sales engines humming
profitably is a cornerstone of all successful compa-
nies. It is especially important in the semiconductor
industry, where the leaders in each product seg-
ment usually take most of the profits. Things often
go wrong when companies feel their market
share is threatened. Frequently, the first response of
both leading and lagging players is to discount
prices. Closer analysis, however, often reveals that
this move needlessly dilutes value that could
have been retained despite competitive pressure.
In semiconductors, discounting is not just
a strategy but an industry norm—a condition that
has prevailed owing to an industry adaptation
to extremely tight innovation cycles. That is, the
larger issue underlying habitual discounting
in semiconductors is the acceptance on the part of
semiconductor companies that a dynamic built
into the industry as a whole will always drive prices
down. Companies therefore often fail to address
and can even contribute to the shortfall by expect-
ing nothing better. However, companies can
begin to reverse this trend by revisiting the value
proposition they are offering to customers,
specifically to determine whether the true, unique
value of their products is being communicated.
This is where the journey to value selling begins.
The value-selling approach
We have devoted considerable research and drawn
upon extensive engagement experience to create
18. 16 McKinsey on Semiconductors Autumn 2011
an approach to sales by which companies can
better understand their existing value propositions,
identify where these are failing them, and make
the transition to a customer-centric approach that
immediately demonstrates the full value of
their products to customers. A value-selling
approach allows companies to pursue and capture
margin according to the end-to-end value provided
to the customer, as opposed to the traditional
cost-plus approach. Even in the semiconductor
industry—with its engineering culture that
is particularly comfortable with cost-plus pricing—
most client executives have been pleasantly
surprised to see the top-line impact possible from
value selling, especially as it requires no changes
to the existing product portfolio. In the first year,
value selling can add 2 to 4 percent to average
selling prices, a number that can grow to
6 to 8 percent in the second year if customer
insights are fed back into the design of
new products.
Our original findings were based on practical
experience. We interviewed dozens of one
company’s end customers, asking each of them to
allocate 100 points among the different factors
in their process for making purchasing decisions.
The great majority of those surveyed identified
price as the most important factor in their final
decision. Deeper analysis of those responses
revealed, however, that some groups of customers
actually cared as much or more about features
than about price. It turned out that some customers
were knowingly paying more for the client’s
product than they would have for the next-best
alternative. The discovery of this partly
hidden preferential-buying behavior led to
a conclusion that a sales approach that
communicates a holistic view of value could
positively affect the way that customers
think about purchases.
The transition to a value-based selling approach
involves a fundamental change in existing
sales approaches and marketing messages. The
constituent tactical shifts of the approach
are made based on deep-structure interviews with
customers, as well as by working to identify
the value of the company’s offerings. A company
adopting this approach will also need to
construct a set of new tools for the sales force and
build value-selling capabilities in the workforce.
Last, it must develop metrics to track sales-force
performance over time. A formal transformation
program is necessary to sustain the transition; the
good news is that the scope can be either
narrow or broad, focusing on a single product family
or region, or encompassing a program that
repositions a company’s entire product portfolio.
Identifying value to the customer
Value selling requires, at its base, an under-
standing of the true value of the product to the
customer. Most customers think of value as
apparent value—that is, the benefit they receive
directly from a purchase. They rarely consider
indirect benefits, such as the positive or negative
impact of switching costs, or the infrastructure
savings that come from sticking to current product
lines. In addition, a mind-set has been observed
in both companies and their customers by which
current prices are expected wholly to deter-
mine future prices. Accordingly, unless the product
changes, the price should rise only in line
with inflation or if particular input costs increase.
This generally leads companies to neglect the
question of how the value of their products evolves
over time.
In all, there are four components to implementing
a value-selling approach: developing a new
approach to purchasing criteria, building a new set
of tools to support the sales teams in the field,
19. 17Getting customers to say, ‘The price is right!’
developing training to educate sales teams on the
new sales approach, and developing metrics
and systems to track the sales force’s performance
over time.
1. Uncovering customer purchasing criteria
The first step in the new direction is a diagnostic
analysis aimed at uncovering the key criteria
that customers use to make purchasing decisions.
This may seem counterintuitive, as company
sales and marketing teams would be expected to
know these criteria already. Typically, top
salespeople do know exactly what matters to
their customers, but their less skilled
colleagues may not; best practices are not
always shared or institutionalized in
a form that is useful to less experienced account
managers. We begin with interviews to
determine the elements that matter most to
customers and to understand how these
are prioritized. The exercise is highly qualitative,
requiring dozens of interviews throughout
the customer base.
The diagnostic yields three results: a list of
factors that are most important to customers, a
ranking of those factors in order of importance,
and, critically, an assessment of company
performance against those value drivers in any
particular product category.
To go beyond articulated preferences, we
conduct a conjoint analysis to quantify value
drivers at both the individual product
level and the broader company level. Typical
findings from diagnostics of the sources
of value at the product level have included
these attributes:
• Superior performance with regard to power
efficiency, speed, graphic resolution, and so on
• Ease of use and ability to design in, without
many additional modifications
• Advantaged size and packaging to enable
greater bill-of-materials integration
• A better set of built-in capabilities to reduce
the need for additional discrete peripherals,
board design, or programming
At the company level, we have found that cus-
tomers derive value from the following elements:
• The reputation of the provider—for example,
the value of saying “powered by x” or
“includes y processor”
• Strong development and technical-
support capabilities
• Advantaged supply-chain and manufacturing
capabilities, resulting in higher product
availability, better quality assurance, and
superior reliability
• Developer tools and well-established
code libraries—for instance, a network of
independent software vendors (ISVs)
for microprocessors
In a value-selling pilot, one customer said, “If it
saves me development cost and improves my
time to market by x percent, I am not going to go
for a cheaper device.” This remark neatly
sums up the reasoning behind value selling and
the source of its impact.
2. The value-selling tool: Communicating
quantifiable value to your customer
Once a company understands the real value of
its products to its customers, it must develop
20. 18 McKinsey on Semiconductors Autumn 2011
easy-to-use tools to guide the sales team in its
work. The tools are often mechanisms for
quantifying the impact of relevant value drivers
and simplifying any calculations the sales
team must make. If we consider the example of a
company that can deliver a particular product
faster than a competitor, its sales force must be
able to quantify the value to the client of
faster time to market in the specific marketplace
into which the product is being sold. A calcu-
lation of such specificity can rarely be performed
in an ad hoc manner by a sales team in the field.
The team therefore needs an analytic tool
that can systematically identify the value created
for the customer by the company’s product
versus the next-best alternative. This identifica-
tion will come in the form of an assessment
that clearly captures the incremental value
delivered by the product vis-à-vis competitor
offerings (Exhibit 1). The tool must be easy
to use, and its analytic mechanism dynamically
responsive to the needs of the field. The
salesperson must be able to enter a few essential
variables reflecting the specific interest of any
Exhibit 1 Research shows that customers make clear value
distinctions for many products.
MoSC 2011
Sales and marketing
Exhibit 3 of 4
Source: Customer interviews; McKinsey analysis
Product value vs NBA, $End device Value/NBA
Client product value
Next-best alternative (NBA)
Product 1 3.6x7.55
2.10
Product 4 1.7x13.00
7.50
Product 5 1.2x4.10
3.50
Product 2 2.6x7.52
2.90
Product 6 1.1x6.80
6.20
Product 7 1.0x9.00
9.00
Product 8
Wide range in value
over NBA price
1.0x2.00
2.00
Product 3 1.8x0.90
0.50
21. 19Getting customers to say, ‘The price is right!’
given customer and obtain an assessment of the
value of the product to that customer.
Much technical and experiential expertise will
be incorporated into the value-selling tool,
and this input must be regularly updated with
product, customer, and market data. The
tool’s purpose, however, is a simple and visible
demonstration of the power of the value-
selling approach. This power lies in value
selling’s ability to communicate a quantifiable
value, based on the drivers that are most
important and therefore worth the most to the
customer relative to competitive offerings.
The approach is entirely oriented to allowing the
salesperson to take a customer-centric view.
It enables salespeople to provide in a dynamic
fashion the information needed by a customer’s
line-of-business leader or procurement
officer in order to make the winning case for
their product to his or her boss. The salesperson
thus delivers to the customer a transformed
value proposition. The message becomes “I’m
getting a product worth two to three times
its price,” rather than “I’m getting a product
10 cents cheaper than the next guy is.”
Not all customers are the same: some may share
a consistent set of value drivers, while others
have variable priorities or find different value in
comparable attributes. Good value-selling
tools empower salespeople to identify the types
of customers they are dealing with and
to generate quickly the most convincing value
proposition for them. The sales force can
adapt to different types of customers without
having to improvise value propositions for
each customer visit (Exhibit 2).
Exhibit 2 What customers value differs significantly according to
the segment they fall into.
MoSC 2011
Sales and marketing
Exhibit 4 of 4
A “Sparrow”
Cannot
benefit from
bulk of
what product
can offer
D “Eagle”
Pushes
the limits—
demands
maximum
performance
on all
metrics
C “Hawk”
Values 2 of 3 attributes; primary
concern is performance over
costs; willing to pay a lot of money
B “Crow”
Values only 1of 3 attributes
but willing to pay a premium
for that value
Low
Design
intensity
Functionality
of chip
Low High Low High
High
End-user
characteristics
Dynamic
Static
22. 20 McKinsey on Semiconductors Autumn 2011
3. Building value-selling capabilities in
the sales force
The value-selling approach represents a
new way for the field sales force to relate to its
customers. How can you bring this new
approach to life for them? Here we believe there
are three critical success factors: first, leaders
and the sales force must co-create the solutions;
second, leadership must visibly model the
new approach; and third, salespeople should be
allowed to practice new skills in a safe, non-
customer-facing environment.
Co-creating solutions
The more the experienced members of the sales
force are engaged in the quantification process—
that is, in generating the underlying numbers
that drive the calculations made by the tools the
sales force will use—the more effective
the result. There are two reasons for this: first,
experienced salespeople understand the
dynamics in the marketplace and can use their
expertise to improve the tool, and second,
they must believe in the calculations that drive
the value-selling tools they are being asked
to use.
Modeling the approach
The second success factor is no less important
than the first. The sales force must feel the
inspiration from above. The head of sales
for the company must participate in the effort
and personally emphasize the need for
value selling.
Practicing the approach
Finally, we have found that it is crucial to create
a low-stress environment, in which sales-
people can practice free of exposure to customers
and fears of internal evaluation. Every time
we have conducted such sessions, we have
discovered that recording the salesperson’s cus-
tomer approach and then reviewing it with
them helps prepare them for even the toughest
customers. When they see themselves
working—and improving—in the new way, sales-
people truly internalize the arguments they
must make and embrace the value quantifications
they will put forward.
Early practice often becomes a learning oppor-
tunity, in which the shortcomings of the existing
approach are revealed and understood as such.
Salespeople usually begin by talking in cost-plus
language, rather than adopting a customer
perspective. A salesperson may fail to offer quan-
tified comparisons with competitive products,
even when supporting data are at the ready. The
key is to keep trying: with practice, the new
way of working begins to stick.
Value selling requires, at its base, an understanding of
the true value of the product to the customer
24. 22 McKinsey on Semiconductors Autumn 2011
Since the late 1980s, the Chinese government has
made efforts to build an indigenous semiconductor
industry by providing financial incentives, devel-
oping talent and technology, and crafting alliances
with global players. And though the country has
assumed a central role in the manufacture of many
computing and consumer-electronics products,
its role in the semiconductor sector has remained
surprisingly limited. In the industry value chain,
China has a strong share in only the assembly-and-
test and back-end-manufacturing segments.
Aside from these two (admittedly considerable)
areas, the country is largely missing from
semiconductor league tables.
In fact, today China is primarily a consumer of
semiconductors, rather than a producer of them.
Sri Kaza, Rajat
Mishra, Nick
Santhanam, and
Sid Tandon
The challenge of China
The country’s semiconductor trade association
published a report in March 2011 that estimated
that the Chinese semiconductor market
accounted for fully 33 percent of global supply. The
share of those chips used in domestic products
accounts for 15 percent of the global semiconductor
market. The remaining share is installed in
a wide range of export goods. Furthermore, our
research indicates that Chinese companies
influence the design and other elements of just 1 to
2 percent of finished chips.
It would be logical to expect that a country
of China’s size would be a leading stakeholder in
discussions about technology standards and
the designs for next-generation platforms, but that
is not the case. Despite consuming 33 percent
As barriers to Chinese competition weaken, local and foreign semiconductor players
must consider issues such as intellectual property and knowledge transfer to fully
capture opportunities in this important market.
25. 23
Exhibit 1
Chinese foundry
TSMC 0.13 1
VIS 0.20 0
Dongbu
Electronics
0.25 0
I-NEC 0.25 0
ALTIS 0.18 0
Grace 0.18 0
SilTerra 0.18 0
SSMC 0.18 0
HeJian 0.25 1
China Resources 0.50 3
Jazz 0.25 1
Tower
Semiconductor
0.18 1
FAB 0.35 4
EPISIL 0.50 3
ASMC 0.50 2
SMIC 0.18 0
UMC 0.15 1
Chartered
Semiconductor
0.18
0.045
0.09
0.09
0.13
0.13
0.13
0.09
0.13
0.15
0.25
0.13
0.13
0.13
0.25
0.25
0.09
0.045
0.045 0
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
3
2
1
Company 2009 nodes, microns
Mainstream1 Leading edge 6”
8
2
2
2
1
1
1
1
1
1
1
2
2
2
1
5
7
5
8” 12”
Number of fabs
Chinese foundries lack leading-edge technologies
because of export controls.
MoSC 2011
China
Exhibit 1of 5
1“Mainstream” includes nodes utilized in more than 50% of the foundry’s total capacity.
Source: iSuppli, H1 2009; World Fab Watch 2009; Wassenaar Arrangement Web site, Category 3 list;
Semiconductor Equipment and Materials International; Taipei Times
• Until 2010, controls prevented the
export of 90nm manufacturing technology
to China
• Until 2010, Taiwan restricted the export
of 180nm manufacturing technology
to China
• In 2010, the Wassenaar Arrangement
was updated: controls now prevent 65nm
manufacturing technology from being
exported to China
• In 2010, Taiwan signed an agreement
with China (the Export Promotion and
Cooperation Agreement) that allows
the export of manufacturing technology that
is 2 generations behind leading edge
The 1996 Wassenaar Arrangement
puts export controls on manufacturing
technology . . .
. . . which has prevented Chinese
companies from accessing
65-nanometer (nm) node technology
of the global market for semiconductors, Chinese
companies claim less than 4 percent of global
revenue in the lucrative segments of semiconductor
design and front-end manufacturing.
There are four reasons for this state of affairs.
First, China exerts little influence on
semiconductor design and selection in major
product categories such as mobile phones
and laptop computers. The majority of design
decisions for these goods are made by
global champions—such as Nokia, Acer, and
Apple—in their home countries, at the
headquarters level.
26. 24 McKinsey on Semiconductors Autumn 2011
Second, the home countries of major semicon-
ductor companies ban the export of leading-edge
manufacturing technologies to China. Both
the United States and the island of Taiwan prohibit
the export of equipment used to manufacture
sub-65-nanometer process technologies, which
leaves mainland Chinese foundries two
generations behind the current 32-nanometer
standard (Exhibit 1).
Third, concentrated clusters of semiconductor
excellence failed to fully develop in China. Instead
of focusing investments on one location,
as did the island of Taiwan with Hsinchu Science
Park, the Chinese government made invest-
ments in multiple provinces, setting up semicon-
ductor fabrication plants as far north as Jilin
and Dalian, as far south as Shenzhen, and as far
west as Chengdu. In all, fabs capable of
producing more than 1,000 wafers per month are
spread across 19 cities. Because the industry
was so fragmented, government support did not
lead to the formation of a vibrant semiconductor
ecosystem in any single location.
Fourth, and perhaps most important, foreign players
own most of the intellectual property throughout
the semiconductor value chain. Applied Materials,
for instance, dominates manufacturing equip-
ment, while Intel, Nvidia, and Qualcomm control
key parts of integrated-chip design for
microprocessors, video cards, and mobile handsets,
Exhibit 2 Lack of intellectual property and know-how
remains the only barrier to increased competition
from Chinese semiconductor players.
MoSC 2011
China
Exhibit 2 of 5
Source: iSuppli; Databeans
Barriers are weakening
End-system design in
China, for China
Declining share of
leading-edge nodes
Renewed focus on clusters
of excellence
Details
Increasing end-user consumption in China is likely to
drive local system design
• ZTE became the No. 5 player in mobile handsets in 2010
• Huawei is a top 3 player in all major telecommunications-equipment segments
• Lenovo is the No. 3 PC vendor in the world
The effect of Moore’s Law on the global semiconductor market is declining
• Leading-edge nodes now represent 14% of total demand for logic chips and
microcomponents, making access to manufacturing technology less important
• Several large segments, such as analog integrated circuits ($42 billion in 2010) and
microcontrollers ($18 billion), are using older technology
Several Chinese cities are beginning to attain critical mass as
clusters of excellence
• Shenzhen, Chengdu, and Dalian have made significant progress
• Texas Instruments and Freescale Semiconductor set up manufacturing
plants in Chengdu
• Intel set up a 90-nanometer fab in Dalian
Issues related to intellectual-property know-how
are the only major roadblock
27. 25The challenge of China
respectively. Owning the intellectual property
means these foreign players also garner the lion’s
share of revenues. In the front-end-manufacturing
segment, non-Chinese players (for example,
Samsung, Intel, and Hynix) earn 96.3 percent
of all revenues. In design, foreign players
earn 96.1 percent of revenues. Even in the silicon
segment, 93.0 percent of revenues go to non-
mainland-Chinese companies. China has a decent
share in only two areas, back-end manufac-
turing and assembly and test, where Chinese
companies earn 28.6 percent of total
segment revenues.
Taken together, these four hurdles have made it
difficult for Chinese semiconductor players
to compete in the last decade. However, three of
those four barriers are now weakening, and with
recent events likely to serve as a tipping point, we
believe the lack of intellectual property and
know-how is the remaining impediment to Chinese
semiconductor players’ progress. This
portends significant shifts in the international
semiconductor situation (Exhibit 2).
The emergence of a Chinese middle class
is creating a domestic industry—one with
export ambitions
The first barrier—the modest influence China
exerts on semiconductor design and selection in
major product categories—is eroding as a
robust domestic market emerges, particularly
because first-time consumers of major
product categories that use semiconductors do not
need leading-edge products. As a consequence,
a substantial “built in China, for China” market is
taking shape. To get a sense of the scale of this
market, consider the following facts: 26 percent of
all automobiles sold in the world in 2010 were
sold in China. Chinese citizens bought 19 percent
of the global PC supply last year and accounted
for 18 percent of LCD-TV sales. In the robust global
market for mobile handsets, the Chinese
commanded 14 percent of unit sales in 2010. And
Chinese companies are leveraging their
domestic scale to sell outside of China, thereby
shaking up league tables further in a number of
industries. Lenovo, for example, is now
the third-largest vendor of PCs in the world. ZTE
became the fifth-largest handset manufac-
turer in the world in 2010. And Huawei has
become a top-three player in all major segments of
the telecommunications-equipment market.
China’s emergence is significantly enabled by a
declining need for ever-increasing processing
speed. As the semiconductor industry moves closer
to the physical limits of silicon, fewer devices
are relying on truly leading-edge technologies. In
fact, leading-edge nodes now represent only
14 percent of total demand for logic chips and
microcomponents. There is, consequently, generally
less pressure to have state-of-the-art manu-
facturing technology (Exhibit 3). This opens the
door for Chinese semiconductor players.
Certain segments of the market have found success
using technology that is one or two generations
behind the leading edge. For example, analog inte-
grated circuits and microcontrollers (which
account for $42 billion and $18 billion in revenues,
respectively) are leveraging process technology
that is at least two years old. The proliferation of
devices powered by less-than-cutting-edge
chips means that the playing field for Chinese
semiconductor manufacturers is much
more level than ever before.
Even if consumers in China become less willing to
settle for second-best technology as their
affluence grows, share is unlikely to shift decisively
back to the West. Chinese semiconductor
companies are developing process technologies
28. 26 McKinsey on Semiconductors Autumn 2011
more quickly. SMIC has now achieved the
same two-year development cycles as industry
leaders. Even though the company may be
at a disadvantage due to Western export controls,
it achieved stable output at the 65-nanometer
level in 2010 and is ramping up additional capacity
in 2011. And SMIC’s 65-nanometer fabs are
running at 95 percent capacity, indicating that
there is intense local demand for these chips.
So the emergence of a local market and
the apparently limited effect of Western export
controls mean that the first two barriers to
a significant Chinese presence in all segments of
the semiconductor industry are coming down.
The third barrier is also falling, because clusters of
excellence are finally coming together in China.
Several cities, including Shenzhen, Chengdu, and
Dalian, have developed expertise in the local
workforce, reached a critical mass in number of
fabs, and connected with relevant suppliers
nearby. A sure sign of this evolution is that Texas
Instruments and Freescale Semiconductor
have both opened manufacturing plants in Chengdu,
and Intel has set up a $2.5 billion 90-nanometer
fab in Dalian.
Looking ahead to the coming decade, it is important
to note that China has the world’s most compre-
hensive, well-funded, and ambitious technology-
Exhibit 3 Leading-edge nodes are only a small share of foundry volume.
MoSC 2011
China
Exhibit 3 of 5
Total foundry capacity per node,
300mm equivalent KWPM1
2,000
1,500
1,000
500
0
≤32nm168
327
322
307
78
45nm
65nm
90nm
130nm
180nm
250nm
350nm
≥500nm
2004 2005 2006 2007 2008 2009 2010 20112 20122 20132 20142
1Thousands of wafers per month.
2Estimated.
Source: Gartner, converted from 200mm to 300mm scale with 8” equivalent to 2.25
29. 27The challenge of China
industry policy, and the semiconductor sector
is 1 of the 16 sectors into which stakeholders want
to make significant inroads. The country’s
industrial policies for semiconductors are already
beginning to show results as domestic end-to-
end value chains emerge: for example, in wireless-
communications semiconductors, an end-to-end
value chain has developed among SMIC, HiSilicon,
Huawei, China Mobile, China Unicom, and
China Telecom. Similarly, in wireless systems on
a chip, the domestic value chain consists of
Taiwan Semiconductor Manufacturing Company
(TSMC), Spreadtrum, Huawei, Tianyu, China
Mobile, China Unicom, and local consumers.
With three of the four barriers weakening, the
lack of intellectual property and know-how is the
only significant barrier remaining. While
the Chinese have found many ways to acquire
the intellectual property needed to establish
domestic industries, challenges related to complex-
ity and materials science in semiconductors
are more burdensome than in other fields.
Acquiring intellectual property and know-how will
thus be crucial for Chinese players as long as the
semiconductor sector remains a priority industry
for government development programs. It should
be noted that China has made multiple attempts to
entice foreign players to transfer technology,
for example, licensing Geode microprocessor-design
technology from AMD.
What does this mean for a strategic China
engagement model?
As the Chinese government increases efforts to
develop the industry, it will likely offer more
promising incentives for semiconductor companies
to do business in China. This creates a dilemma:
it will be difficult for foreign companies to compete
from outside the country as their competitors
establish beachheads there, but at the same time,
the Chinese endgame is clearly the transfer
of intellectual property and know-how to allow
Chinese companies to compete globally—
that is, not just to compete for the emerging local
market currently owned by Western players,
but to turn around and challenge Western players
on their own turf.
A similar scenario played out in the mid- to late
2000s, when the Chinese government launched a
major policy initiative to promote the high-
speed-rail industry. Seeing a $50 billion market,
many foreign players, including Kawasaki,
Siemens, Bombardier, and other companies,
expressed interest. In 2004, several joint ventures
were set up between foreign and local rail
companies. While Siemens refused to transfer
intellectual property to its joint-venture
partner without adequate compensation, Kawasaki
and one other player agreed to transfer signifi-
cant intellectual property to their respective
partners on less demanding terms.
As classic game theory would predict, had all three
players sold into China on similar terms, the
$5.2 billion in potential reward would have been
divided among the three equally ($1.7 billion
for each, as seen in the top-left box in Exhibit 4).
Since, however, Kawasaki and one other
player agreed to Chinese terms on intellectual
property but Siemens did not, it seemed
likely that the two companies that shared IP would
split the market equally. But Siemens felt it had
little choice but to set up a joint venture including
intellectual-property transfer to claim its
share of revenues, so another scenario from the
initial game-theory projection played out:
each of the three players divided their share of
the $5.2 billion market equally, and then split that
amount with their Chinese partners.
However, at the three-year mark, all the local
joint-venture partners, having carefully
30. 28 McKinsey on Semiconductors Autumn 2011
incorporated key intellectual property from the
foreign players, began launching indepen-
dent products. Since 2007, these products have
attracted $20 billion in orders from
various state-owned enterprises; foreign players
have not won any orders at all.
This cautionary tale is not presented as
definitive proof that the joint-venture structure
is flawed irremediably. Rather, we mean to
suggest that other structures must be energetically
reviewed; companies should consider
options that do not include the transfer of
intellectual property.
For instance, a number of leading multinational
companies have adopted an “innovate with
China” approach, which consists of launching RD
centers in China that focus on developing
technologies for the Chinese market. General
Electric, for example, established a China
RD center that focuses on developing products in
line with local market demand and stated
government priorities, such as rural health care
and sustainable development. Siemens has
a similar center working on LED lighting products
and low-cost medical equipment. Each product
from these centers is tailored to the Chinese market
and could potentially be sold in other developing
Exhibit 4 Classic game theory can be used to predict potential outcomes
of partnerships in Chinese high-speed rail.
MoSC 2011
China
Exhibit 4 of 5
Source: New York Times; Financial Times
Once Kawasaki and another foreign player agreed to joint ventures, Siemens had to do so as well
Potential rewards, 2004–06: total orders of $5.2 billion
All 3 players get equal
share of $5.2 billion
market and keep 100%
of revenue
Player with joint venture
is preferred supplier and
gets entire market;
$5.2 billion revenue is split
with joint-venture partner
All 3 players get equal
share of $5.2 billion
market, but they must
split revenue with
joint-venture partner
Player with joint venture
is preferred supplier and
gets entire market;
$5.2 billion revenue is split
with joint-venture partner
Sell into Chinese market Local joint venture with
intellectual-property transfer
Actions of Siemens
Actions of
foreign
players such
as Kawasaki
Sell into
Chinese market
Local joint
venture with
intellectual-
property transfer
1.7
1.7
0.9
0.9
0
2.6
2.6
0
$ billion for
Siemens
$ billion for
foreign players
such as
Kawasaki
Win-win
Stable state
31. 29The challenge of China
markets. This approach serves to limit the exposure
to intellectual-property risks to technologies or
products developed in China (Exhibit 5).
More broadly, there are a few simple steps that
foreign players can take to boost their chances of
success in the Chinese market. Keys include
developing a go-to-market approach that addresses
the problems of Chinese customers, nurturing
strong relationships with large state-owned enter-
prises, and presenting an innovative in-channel
model to take advantage of unique characteristics
of the market.
Four strategic questions to consider
Until now, foreign players have focused on
protecting their intellectual property and know-
how by selling finished chips into China.
One common tactical approach is known as price
customization; companies offer special
product numbers and packaging, and although
product performance is slightly lower, the
goods cost less. While this approach meets basic
market requirements, it creates an opportunity
for local players; they can add features and
differentiate themselves significantly. To head off
that threat, many foreign semiconductor players
Exhibit 5 Reviewing joint-venture structures can help avoid the prisoner’s
dilemma, as in this electric-vehicles scenario.
MoSC 2011
China
Exhibit 5 of 5
Potential rewards, 2015–20: total potential orders of $4.5 billion1
All 3 players get equal
share of $4.5 billion
market and keep 100%
of revenue
Player with joint venture
is preferred supplier and
gets entire market;
$4.5 billion revenue is split
with joint-venture partner
All 3 players get equal
share of $4.5 billion
market, but they must
split revenue with
joint-venture partner
Is innovating in China, for China
a better alternative?
Player with joint venture
is preferred supplier and
gets entire market;
$4.5 billion revenue is split
with joint-venture partner
Sell into Chinese market Local joint venture with
intellectual-property transfer
Actions of semiconductor company A
Actions of
semiconductor
companies B/C
Sell into Chinese
market
Local joint venture
with intellectual-
property transfer
1.5
1.5
0.7
0.7
0
2.3
2.3
0
$ billion for
semiconductor
company A
$ billion for
semiconductor
companies B
and C
Win-win
Stable state
1The Chinese government set a goal of 5 million electric cars in China by 2020; $900 in semiconductor content
per hybrid or electric car.
Source: New York Times; Financial Times
[[Using
bars t
indica
$ billio
Ok? Is
correc
32. 30 McKinsey on Semiconductors Autumn 2011
A combination of policies designed to enable
large, next-generation end-use markets for
semiconductors, together with procurement policies
meant to drive indigenous innovation, is likely
to create a strategic dilemma for semiconductor
companies looking to sell into China.
China has launched ambitious policy initiatives
to develop large domestic markets for specific next-
generation technologies: cloud computing, the
“Internet of Things,” and hybrid and electric vehicles.
These three markets combined represent tens
of billions of dollars of market opportunity in China
for semiconductor companies.
The Chinese government is also increasingly
emphasizing indigenous innovation in government
procurement programs in order to reduce the
country’s dependence on foreign technology. In
November 2009, several Chinese government
agencies announced six categories of products that
would be directly affected: computer and
application devices, communication products
(thought to include mobile phones), modern-
ized office equipment, software, “new energy and
equipment,” and energy-efficient products.
China’s 12th five-year plan also reinforces the drive
to promote domestic innovation in these areas.
Taken together, these policies and a number of
stimulus programs may have significant implications
for the semiconductor industry. These next-
generation technologies and categories of products
Indigenous innovation and next-generation
markets for semiconductors
are expected to be growth drivers for Western
semiconductor players in the decade ahead. There
is a real potential for Chinese companies to
emerge in these areas, as current players have
not established clear leadership positions
in these applications.
China has not yet tied the indigenous-innovation
policy to its policies for these next-generation
markets. But there is a real possibility that it will.
And any move in that direction would create a
strategic dilemma for semiconductor players, which
are, frankly, counting on driving significant future
growth from these three areas (exhibit).
Simply responding to the challenge of establishing
a presence in these areas by creating individual
initiatives will not be sufficient. This is a matter that
should rise to the highest strategic level for
any company that wishes to be a player in these
markets. A good place to start would be to
understand the implications of potential government
and competitor moves, and to develop a response
that will accommodate each.
33. 31The challenge of China
have begun designing products for China, in China,
yet they remain wary of the risks of transferring
intellectual property and know-how.
From a strategic perspective, there are four key
questions that semiconductor companies
must answer to successfully address the oppor-
tunity in the Chinese market.
The first question concerns the engagement
strategy for intellectual property and know-how.
Simply put, what is the best way to use intellectual
property in China? Two common strategies
are to sell into China while keeping intellectual
property in-house and to launch a joint ven-
ture with an agreed-upon transfer of intellectual
property. However, several other options exist.
Companies could launch indigenous development
centers in China, which will develop key
technologies for the unproven, next-generation
markets likely to take off should they become
widely adopted in China. Another option is for
Exhibit The semiconductor industry may face challenges related to
intellectual property for next-generation applications.
MoSC 2011
China sidebar
Exhibit 1 of 1
The indigenous-innovation policy . . .
• An “indigenous innovation catalog”
of domestically developed technologies
was created: approved products are
to be given preferential treatment in state
procurement
• Initial focus areas include 6 high-tech
industries: targets include computing,
networking, and energy efficiency
• The goal is to move from “made
in China” to “innovated in China”:
the country wants to reduce its dependence
on foreign technology to 30% from its
current level of 50%
. . . could have significant
implications when applied to large,
next-generation markets
• Cloud computing
– Policy initiative launched in 2010
– Trials will take place in 5 cities
– Key goals include developing
core technologies and
formulating standards
• “Internet of Things”
– Policy initiative launched in 2011
– The goal is to develop domestic
leaders in the industrial value chain
– Stakeholders also seek to provide
support to develop standards
• Hybrid and electric vehicles
– Policy initiative launched in 2010
– The aim is to make China a leader
by 2020
– The country seeks to develop
2–3 companies as global leaders in
key technology areas
This may result
in a strategic
dilemma for the
semiconductor
industry
35. 33
Ulrich Naeher,
Sakae Suzuki, and
Bill Wiseman
The evolution of business models
in a disrupted value chain
The progress predicted by Moore’s Law has slowed in recent years. Players across
the semiconductor value chain must adjust their approaches to compete as the industry
continues to evolve.
Over the past decade, the growing importance of
specialization and scale in semiconductors has
led to a breakup of the value chain and the establish-
ment of a “winner takes all” dynamic in many
market segments, as noted in “Creating value in the
semiconductor industry” (p. 5). Scale has become
essential, as technical evolution in line with
Moore’s Law requires larger and larger investments
in RD each year. Specifically, the pursuit of
smaller gate sizes, larger wafers, and competitive
scale has resulted in an increase of about
20 percent in investment per year in leading-edge
technology nodes. As a result, only a handful
of companies—such as Intel, Samsung, and Taiwan
Semiconductor Manufacturing Company (TSMC)—
can keep up in the technology race.
Design costs, measured on a project basis, have
exploded as well, resulting in a reduction of
new designs (Exhibit 1). It is no surprise that Intel
stands out as the winner in microprocessor
units (MPUs), Texas Instruments in diversified
integrated device manufacturers (IDMs),
and TSMC in foundry; Samsung and Toshiba are
arguably the winners in memory. All other
players, including most IDMs, net out with either
zero or negative cumulative economic profits
from 1996 to 2009. Across the industry, semi-
conductor players destroyed a combined
$140 billion in value.
However, not every segment conforms to this
Darwinian model; fabless players and segments
36. 34 McKinsey on Semiconductors Autumn 2011
such as analog IDMs are two examples of busi-
nesses that are less ruthlessly competitive. In fact,
the progress predicted by Moore’s Law has
slowed in many segments of the semiconductor
industry. Given this context, we examined
ways in which the semiconductor value chain
might evolve and explored how current players
might adapt in order to compete.
We begin by looking at the changes occurring
in the fabless segment. Next, we turn to
front-end fabrication, the segment that drove the
technical developments that enabled the kind
of advances predicted by Moore’s Law. Finally, we
address back-end fabrication, where the
miniaturization race seems to be shifting the
assembly-and-test segment.
Fabless design: Players adopt a range of
successful models
Over the last decade, fabless players have continued
to gain ground, outpacing IDMs and claiming
more than 20 percent of the market. Despite some
scale in high-end design, there is no overarch-
ing winner-takes-all dynamic in this corner of the
industry. In fact, as noted in “Creating value
in the semiconductor industry” (p. 5), there are far
more fabless companies generating economic
profits than there are profit-generating companies
in manufacturing-related business domains.
At first glance, one might conclude that fabless
players create value because they require
less capital investment. However, we find these
companies win by establishing dominance in
specific applications rather than across applica-
Exhibit 1 There are fewer design starts, particularly on
newer advanced technology.
MoSC 2011
Moore’s Law
Exhibit 1 of 5
1Integrated circuit.
2Application-specific standard product.
3Application-specific integrated circuit.
Source: Global Semiconductor Alliance; Morgan Stanley; McKinsey analysis
Number of logic IC1 design starts
12,000
10,000
8,000
6,000
4,000
2,000
0
ASSP2
ASIC3
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
37. 35The evolution of business models in a disrupted value chain
tions. Overall, three distinct business models have
succeeded in the fabless space: innovators,
fast followers, and mature-market attackers.
The innovator model is exemplified by leading
players such as Qualcomm. These companies
invest in continuous innovation for new applications,
and they constantly expand their core intellectual
property. Their efforts focus on unmet needs
in the marketplace that come with large potential
demand, and their explicit aim is to provide
targeted semiconductors at the scale required to
recoup RD costs.
But being first to the market is not a must for fabless
players. Broadcom is a good example of a fast
follower. Instead of gambling on untested market
potential, fast followers pick large, rapidly
growing markets and quickly develop intellectual
property to enter certain segments. They position
themselves as presenting an integrated solution
that is a lower-cost alternative to the market leader,
with a streamlined business structure.
The third model, the mature-market attacker, is
best illustrated by MediaTek. It may appear quite
similar to a fast follower at first glance. However,
such companies wait until an application area has
reached significant global volume before entering
the fray. At that point, they attack the market with
a simplified value-for-money product offering.
Execution excellence—that is, efficient development
and speedy production—is crucial for these
players. Other companies in this category include
Monolithic Power Systems, Richtek Technology,
MStar Semiconductor, and RDA Microelectronics.
With the ongoing commoditization of manufac-
turing services and better access to leading-edge
intellectual property, the fabless industry will
profit from its focused business system. We expect
these players to dominate more and more
successful applications, especially in consumer
electronics and some areas of IT. IDMs and
even established microcontroller and micropro-
cessor players will continue to cede ground to
fabless competitors. The compound annual growth
rate of the fabless segment, which currently
outperforms the overall semiconductor industry by
more than 5 percentage points a year, seems
sustainable over the longer term.
A more interesting model that may be reemerging
is that of the integrated original-equipment
manufacturer (OEM). In the past 30 years, many
OEMs, such as Motorola and Hewlett-Packard,
divested their semiconductor arms due to the high
capital intensity of these businesses and the
need for scale. Today, a new generation of OEMs
that are tied neither to in-house process
technologies nor to software development are
taking more ownership of integrated-circuit
design. Apple and Google may indicate the emer-
gence of a larger trend, in which we see that
the intellectual property for functional design may
not belong fully to the chip maker but to a new
kind of integrated OEM. With valuable functional
designs in hand, such players may in-source
or outsource chip design based on cost. Companies
such as Apple and Google have sufficient scale
and capability to become fabless for both the box
and the chips. Because these OEMs tend to be
market leaders, they can compete with innovator
and fast-follower companies for share in
the overall profit pool. Of course, OEMs without
the scale or skills will continue to rely on mature-
market attackers to sustain their businesses.
Front-end manufacturing: What are the limits
of vertical disintegration?
Over the past several years, many semiconductor
companies have decided to go “fab lite,” or step out
38. 36 McKinsey on Semiconductors Autumn 2011
of some aspects of the capital-intensive, leading-
edge front-end technology-development and
fabrication part of the value chain. Nearly all IDMs
have outsourced some of their production to
foundries. Even the Japanese, who are known for
their reluctance to give up in-house capabilities,
are going asset light, at least in part. Examples
include Fujitsu, Renesas Electronics, and
Toshiba. As a result, the foundry business has
surged over the last decade, outperforming
IDMs by an average of about 5 percentage points
each year (Exhibit 2).
In the longer term, the foundry business has
evolved over the last 20 years. Although it earlier
competed on factor cost advantages, produc-
tivity gains, and operational excellence, it now
depends on true technology leadership,
scale advantages, and a superior ecosystem for
product design. Modern foundries can
provide every type of support: for example,
developing intellectual property, offering
photomasks, and offering access to networks of
third-party design centers. Services even
include competence in testing and packaging.
More recently, foundries have started to
offer 3D expertise, interposers, and back-end
integration as a means of differentiating
themselves from competitors.
As noted above, in the early days, the foundry
model generated profits primarily through
low costs. Analysis of manufacturing costs that
pitted European IDMs against Taiwanese
foundries in the 1990s indicated that the cost
advantages of the foundries were close to
50 percent. By the mid-2000s, leading foundries’
process technology reached parity with other
leading-edge players in standard complementary-
metal-oxide-semiconductor (CMOS) technologies.
By the end of the 2000s, foundries became
the core of new technology clusters. They no longer
had to compete on price.
Despite the fanfare, foundry volumes occupy only
20 percent of current manufacturing capacity
Exhibit 2 Fabless and foundry businesses have grown above the industry
average, whereas IDMs have grown below it.
MoSC 2011
Moore’s Law
Exhibit 2 of 5
1Integrated device manufacturer.
Source: iSuppli; IC Insights; Gartner
Revenue growth by value-chain slices,
% compound annual growth rate, 2005–10
4
Overall
3
IDM1
9
Foundry
11
Fabless
39. 37The evolution of business models in a disrupted value chain
(Exhibit 3). With the expertise that many foundries
currently possess, when will this business model
truly take off? The reality is that there probably will
not be any great jump in market share.
Growth in the foundry business has rested on
three pillars: first, leading-edge fabless companies
such as Qualcomm and Nvidia rely on foundries
to produce their designs, including hot products
such as application-specific integrated circuits
(ASICs) and application-specific standard products
(ASSPs), all of which are sold into the global
semiconductor market. A second pillar of growth
has come as a result of IDMs looking to go
fab lite; examples include NXP, Texas Instruments,
Freescale, Fujitsu, and Renesas. The third
growth driver has been increasing share among
existing customers due to foundries’ ability
to produce chips for cutting-edge and trailing
products at a lower cost.
In our market model, we expect most new ASIC or
ASSP capacity to be built within the foundry
ecosystem. In addition, all the new leading-edge
capacity for nonmemory applications will end
up at foundries or Intel. Nevertheless, we assume it
will be difficult for foundries to gain share at the
lagging edge of the chip market because IDMs are
producing them based on sunk-cost economic
models. It will be equally challenging for them
to move in on the specialty technology businesses
of IDMs, which also thrive due to depreciated
assets (and which display relatively low portability
across fabs precisely because of the level
of specialization in these products). Given these
assumptions, the slowdown of Moore’s Law
node migration and the fact that most IDMs have
already turned asset light will impede foundry
growth. In total, we expect the segment to
grow 5 to 10 percent a year, rather than the 10 to
15 percent it grew in years past.
Exhibit 3 Despite early fanfare, foundry volumes
remain low relative to IDMs.
MoSC 2011
Moore’s Law
Exhibit 3 of 5
1Integrated device manufacturers.
Source: iSuppli, Q4 2010; World Semiconductor Trade Statistics, Q4 2010; McKinsey analysis
Semiconductor fab capacity for foundries and IDMs,1
%, total in million 8-inch-equivalent capacity per month
Foundry capacity share
IDM capacity share
100% =
2000
91
9
9.4
2002
88
12
8.9
2004
85
15
10.3
2006
85
15
12.6
2008
83
17
14.2
2010
80
20
14.6
40. 38 McKinsey on Semiconductors Autumn 2011
While foundry demand may be growing less
quickly, a great deal of capacity is coming online.
As a consequence, we expect leading foundry
players such as TSMC, Samsung, Global Foundries,
and United Microelectronics Corporation to
compete for customers more aggressively than they
have in the past, as capital expenditure and
process-technology development costs skyrocket.
Second-tier players from China and Malaysia will
also try to operate at capacity. In addition,
Japanese IDMs might give away surplus capacity
to potential customers at “cash cost,” hurting
the foundries’ price points. All in all, price compe-
tition in this sector will likely intensify.
Exhibit 4 Four leading OSAT players adjusted their business
models and are generating economic profits.
MoSC 2011
Moore’s Law
Exhibit 4 of 5
1Calculated as (return on invested capital – weighted average cost of capital) × invested capital.
2Outsourced semiconductor assembly and test; the 4 players referenced in the chart are Amkor Technology (Amkor), STATS
ChipPAC (Stats Chip), Siliconware Precision Industries Co. Ltd. (SPIL), and ASE Global (ASE).
Source: McKinsey Corporate Performance Analysis Tool
Economic-profit generation1 for the top 4 OSAT2 players, $ billion
Value-destroying phase Value-generating phase
1,600
–1,600
1,400
–1,400
1,200
–1,200
1,000
–1,000
800
–800
600
–600
400
–400
200
–200
–175
–131
–393
783
755
167
11
0
2004 2008 20092007200620052003
Amkor –75
ASE –61
SPIL 0
Stats –39
Chip
Amkor –125
ASE 38
SPIL 58
Stats –101
Chip
Amkor –147
ASE –266
SPIL 122
Stats –103
Chip
Amkor 138
ASE 368
SPIL 239
Stats 39
Chip
Amkor 131
ASE 272
SPIL 348
Stats 4
Chip Amkor 70
ASE 21
SPIL 119
Stats –42
Chip
Amkor 42
ASE –57
SPIL 145
Stats –119
Chip
41. 39The evolution of business models in a disrupted value chain
More recently, the earthquake and tsunami in
Japan brought the need for supply-chain
diversification to the forefront. Specialization and
geographic concentration, which helped drive
success in foundries in the past, are now becoming
risks. Will foundry companies be able to
provide risk diversification from natural or man-
made disasters? Will OEMs be willing to bear
the infrastructure costs associated with having
multiple suppliers on fragmented campuses
manufacturing interchangeable and commoditized
technologies? The answers to these questions
are unclear at this time. However, continued
business-model innovation is needed to enable
multisourcing with minimal cost impact,
if not further cost reduction. If this evolution can
be achieved, it will likely drive continued
disintegration in the semiconductor value chain.
All in all, it does not look as if the foundries’
current 20 percent market share will grow
appreciably anytime soon. Indeed, in addition to
interfoundry competition, these players
are quite likely to face competition from other
players along the value chain such as Intel
and Samsung.
Given these facts, and the implications of
deceleration with regard to Moore’s Law, how will
foundries capture a fair, if not disproportionate,
share of the profit pool? From the other side, how
will customers capture more value from the
foundries? Naturally, the big foundries would favor
fewer foundry players. At the same time,
OEMs, IDMs, and fabless players would prefer
to have multiple leading-edge foundries.
Investment for capacity, partnerships, alliances,
and distribution of orders across foundries
over the next three to five years will be crucial in
determining the competitive dynamics of
the industry.
Back-end manufacturing:
The race for miniaturization brings success to
OSAT players
Chip packaging has shifted to an outsourcing
model more quickly and more extensively than
front-end processes have. In fact, many
expect that the outsourced share of this segment
could reach 50 percent of the market by 2013.
In the not-too-distant past, outsourced semicon-
ductor assembly and test (OSAT) companies
were regarded as low-end, commoditized service
businesses, and the competitive dynamics
of the business were driven by price competition.
This had a negative impact on the economics
of the industry. Between 1996 and 2006, the sector,
cumulatively, delivered no significant economic
profit. However, the same analysis for the
next three years shows a very different result
(Exhibit 4). Part of the OSAT industry has
undergone a transformation, and there are now
two profitable subsegments: the very profit-
able high-end players and the less successful
mainstream OSAT players.
As the pace of innovation slows in the front-end
segment of the semiconductor market, the
pressure on back-end companies is increasing;
these players are expected to offer sophisti-
cation and technical differentiation in a bid to
increase chip performance. Technological
differentiation will continue to drive the two-tier
market structure: there will be oligopolistic
high-end players and commoditized mainstream
players. The OSAT industry can stay profitable
at the high end as long as the top players have the
technical skills required to differentiate
the degree to which the chips they receive from
foundries can be tuned to the needs of
different products—and if they are able to avoid
price wars.
42. 40 McKinsey on Semiconductors Autumn 2011
Four leading OSAT players have redefined their
businesses models successfully and are generating
economic profits. Amkor Technology, STATS
ChipPAC, Siliconware Precision Industries Co. Ltd.,
and ASE Global are the four high-end OSAT
companies, and each has significantly improved its
profitability since 2006, leaving aside some
turmoil caused by the Lehman Brothers collapse
and resulting economic downturn.
These companies successfully migrated from
value destruction to value creation by focusing on
improved capital productivity through careful
management of investments and by introducing
more sophisticated pricing models. Further-
more, they invested in advanced packaging tech-
nologies and improved miniaturization
technologies, such as ball grid array (BGA) and
flip-chip BGA, and shifted their product
portfolio to those categories. The top four OSAT
companies account for 80 to 90 percent of all
outsourced substrate-based packaging services
(Exhibit 5). On the other hand, lead-frame
packaging services have become essentially a
commoditized market. Technology-based
differentiation allows certain players to access
more specialized markets. In those narrower
niches, pricing pressures are much lower than they
are in the more commoditized packaging
segments. Companies thus want to be in the sub-
strate rather than the lead-frame packaging-
services game.
These four companies avoided the vicious
price competition of the early 2000s, and they
also improved their cost position, reducing
capital expenditure by planning capacity more
carefully and by avoiding unnecessary
capacity buildup. One way they have done this
is by maintaining leadership in technology,
Exhibit 5 Revenues can be broken down by package type.
MoSC 2011
Moore’s Law
Exhibit 5 of 5
1Outsourced semiconductor assembly and test; the 4 players referenced in the chart are Amkor Technology (Amkor),
STATS ChipPAC (Stats Chip), Siliconware Precision Industries Co. Ltd. (SPIL), and ASE Global (ASE).
2Figures may not add up to 100% because of rounding.
3Includes bumping.
4Wire-bonded ball grid array.
Source: Company filings; expert interviews; McKinsey analysis
OSAT1 revenue split (among top 4) by package type,2
%, FY 2008
Top 4 players
36 27 23 5
22 17 23 18
5 3 9 3
13 6 7 9
14 10 13 8
10
20
79
65
54
Flip chip3
ASE SPIL Amkor OthersStats Chip
WB BGA4
Lead frame
Testing
Overall
$2.3 billion
$4.5 billion
$9 billion
$4.6 billion
$20 billion